Yes. It seems as if this would be a tremendous step toward the future and would help everybody.
On 1/29/2016 4:21 PM, Joy Nelson wrote:
Agreed Tomas. I would love to see Koha have the ability to be ‘format’ agnostic when it comes to the data.
On Fri, Jan 29, 2016 at 9:14 AM, Tomas Cohen Arazi <tomascohen> wrote:
There are (at least) two options:
– Make Koha handle more metadata formats (for authority records /
thesaurus) and have the tools extract the needed heading for the
– Make Koha handle importing SKOS-core packages into its authority database.
Those could work IMO.
2016-01-29 12:11 GMT-03:00 James Weinheimer <weinheimer.jim.l>:
> On 1/27/2016 11:19 PM, Craig Butosi wrote:
>> I’ve been curious for a while now if anyone has successfully
>> bulk-uploaded into Koha (3.20 or newer) one of the free LC Name authority
>> database files available through the Library of Congress:
>> I wonder if it’s even possible given the file formats in which these
>> files are available (n-triples, rdf/xml, Turtle, etc). It looks like some
>> sort of conversion would be required before uploading into the Koha
>> authorities module. It would be lovely to bulk upload a Name authority file
>> direct from LC rather than Z39.50’ing individual authorities.
> This would be very positive. One of the problems is that the downloads
> available are not MARC, but are either MADS or SKOS. It seems as if
> converting these records to MARC would lose information, but I am not sure.
> To see what this means, look at the record “United States–History–Civil
> War, 1861-1865–Bibliography”
> http://id.loc.gov/authorities/subjects/sh2007100437.html. If you select
> RDF/XML (MADS and SKOS) (
> http://id.loc.gov/authorities/subjects/sh2007100437.rdf) this is what
> would need to be translated into MARC.
> The guidelines for mapping MARC to MADS are at
> http://www.loc.gov/standards/mads/mads-mapping.html, and it says “This
> mapping gives equivalencies between the MARC 21 Authority Format and MADS,
> but is not intended to be a crosswalk that allows for bi-directional
> conversions without some loss of data. Where there is no equivalent MARC 21
> element, this mapping says “no MARC 21 equivalent”. Therefore, I assume
> there are some problems.
> I don’t know if anybody has experience with this, but I don’t think the
> problems would be trivial.